Ceramics International 29 (2003) 1317 www.elsevier.

com/locate/ceramint

A new sintering additive for silicon carbide ceramic

Z.H. Huang*, D.C. Jia, Y. Zhou, Y.G. Liu
Materials Science Department, Harbin Institute of Technology, Harbin 150001, China Received 20 December 2001; received in revised form 25 January 2002; accepted 11 March 2002

Abstract Silicon carbide (SiC) ceramic with nanoscale silicon carbide powder and oxide (MgO+Al2O3+Y2O3) or nitrate salt [Mg(NO3)2+Al(NO3)3+Y(NO3)3] additives was hot-pressed at 1800  C, 25 MPa, for 30 min. All the samples had ne, equiaxed microstructures with an average grain size less than 1 mm. The addition of nanoscale SiC did not eectively improve the relative density. Whether nano-SiC was added or not, the samples with nitrate salt additives had a higher relative density than their counterparts with oxide additives, and the maximum values were as high as 98.1 and 97.7%, respectively. Crystalline YAG phase was detected by XRD only in samples with nitrate salt additives, and the major phases in the oxide additive samples were the same 3C and 4H polytypes as those in the starting powders. Compared with the previous study, the elemental distribution of Y, Al, and Mg was more homogeneous in the oxide additive samples. # 2002 Elsevier Science Ltd and Techna S.r.l. All rights reserved.
Keywords: A. Hot pressing; D. Silicon carbide; Sintering additives

1. Introduction Silicon carbide (SiC) is a promising material for high temperature engineering applications, however, it is dicult to densify without additives because of the covalent nature of SiC bonding and low self-diusion coecient [1]. Boron, aluminum, carbon, or their compounds have been successfully used as sintering aids to fabricate SiC ceramics at about 2000  C [2]. Liquid phase sintering with the addition of MgO and/or Al2O3 and/or Y2O3 has been also investigated at 18502000  C [37]. But the additives used to enhance processing invariably become weak secondary phase in the nal ceramic [810], which usually lower its mechanical properties at high temperature. This detrimental eect infers that the smallest fraction of additives is desirable. In addition, the eectiveness of the additives greatly depends on the homogeneity of their distribution [11]. Due to lower sintering temperature than normal powders, ultrane powders have been used as sintering agents [12]. Also, when adopted as sintering additives in
* Corresponding author. Tel.: +86-451-6414-291; fax: +86-4516413-922. E-mail address: zhihenghuang@yahoo.com.cn (Z.H. Huang).

silicon nitride (Si3N4) ceramic, nitrate salts are more homogenously distributed than their oxide counterparts during the ball-milling process [13,14]. The main purpose of the present study is to combine the advantages of the above-mentioned additives. We blend nanoscale and submicron SiC powders to improve sinterability, use nitrate salts instead of oxide as sintering aids, and try to minimize the amount of additives to fabricate silicon carbide ceramic.

2. Experimental procedure Commercially available sub-micron powders of b-SiC (Central Iron and Steel Research Institute, Beijing, China) with an average particle size of 0.5 mm and 70 nm nano-SiC (East Nanometer Materials Co., Ltd, Shijiazhuang, China) were used. Nitrate salts Mg(NO3)26H2O (99.0% purity, Beijing Chemical Plant, China), Al(NO3)39H2O (99.0% purity, Beijing Chemical Plant, China) and Y(NO3)36H2O (599% purity, Special Type Chemical Reagent Development Co. of North China) were adopted as sintering additives. To compare the eects of nitrate and oxide additives, MgO (98.0% purity, Minfeng Reagent Plant of Wuxi, China),

0272-8842/02/$22.00 # 2002 Elsevier Science Ltd and Techna S.r.l. All rights reserved. PII: S0272-8842(02)00082-2

14

Z.H. Huang et al. / Ceramics International 29 (2003) 1317

Al2O3 (Beijing Construction Material Research Institute, China) and Y2O3 (99.9% purity) were also used as sintering aids. The composition of dierent samples was referred to Table 1. The amounts of nitrates were calculated based on their oxide forms. All the individual batches were milled in ethanol for 24 h using Si3N4 balls and then dried in air on a hot plate. When nitrate additives were used, the solutions were added to the bottles for ball milling after the nitrate salts had been dissolved in ethanol. The ball-milled and dried powders with nitrate sintering additives were calcined at 500  C for 2 h in a nitrogen-gas atmosphere to decompose the nitrates. Hot-press sintering was performed on an FVHP-R-50 FRET-100 furnace (Fuji Dempa Kogyo Co. Ltd., Japan) at 1800  C, 25 MPa, for 30 min in an argon atmosphere. The density of the specimen was determined by the Archimedes method. The theoretical density was calculated according to the rule of mixtures. The microstructure was observed by scanning electron microscopy (SEM) system (Model X-650, Hitachi). Before SEM observation, the specimen was polished, and then chemical-etched (boiling Murakami reagent) for 30 min. Crystalline phases were determined by X-ray diraction measurements (D/max1-rB, Japan) using CuKa radiation. The distribution homogeneity of elemental yttrium, aluminum, and magnesium in the

sintered samples was examined by electron microprobe (Jeol Superprobe 733, Japan).

3. Results and discussion 3.1. Relative density and microstructure Figs. 1 and 2 show the relative density of the sintered samples. All the samples with oxide additives have a value above 95.8%, and obviously, the addition of nano-SiC particles decreases the relative density, which is true both in the samples with oxide additives and the ones with nitrate salt additives. As the content of nanoSiC increases, the relative density slightly increases, but it is still lower than the samples without nano-SiC particles. The result is consistent with ndings reported by Sciti [15]. The decrease of relative density after nanoSiC powder addition has probably resulted from the larger amount of silica coming from nano-SiC powder, which increases the amount of liquid phase during hotpressing. More liquid phase will aggravate the volatilization problem and result large size grain boundary phase pockets in the nal ceramic. In the samples with nitrate salt additives, the relative density increases almost linearly with the increasing amount of nitrate salts. Fig. 3 compares the eect on

Fig. 1. The relative density of the samples with 9 wt.% (MgO+

Fig. 2. The relative density of the [Mg(NO3)2+Al(NO3)3+Y(NO3)3]

Table 1 Starting composition of SiC+additives Sample S9O S5S9O S10S9O S15S9O S9N S10S9N S10S6N S10S3N Composition SiC+oxide additives SiC+5 vol.% nano-SiC+oxide additives SiC+10 vol.% nano-SiC+oxide additives SiC+15 vol.% nano-SiC+oxide additives SiC+nitrate additives SiC+10 vol.% nano-SiC+nitrate additives SiC+10 vol.% nano-SiC+nitrate additives SiC+10 vol.% nano-SiC+nitrate additives Amount of additives (wt.%) 9 9 9 9 9 9 6 3

The molar ratio of the additive composition was xed at MgO:Al2O3:Y2O3=67:16:17.

Z.H. Huang et al. / Ceramics International 29 (2003) 1317

15

relative density of oxide sintering aids with that of nitrate salts. It shows that the samples with nitrate salt additives have a higher relative density. Note that the sample with only 3 wt.% sintering additives, hot-press sintered at 1800  C for just 30 minutes has a relative density of as high as 86.8%, which is dicult to achieve in the sintering of silicon carbide ceramic. Figs. 4 and 5 present the secondary electron SEM images of the samples. Note that all the samples have ne, equiaxed microstructures with an average grain size less than 1 mm. This is due to the low sintering temperature, lower soaking time, and the addition of nanoSiC particles. Except for S10S6N and S10S3N, which are less densied and have more residual pores, all the samples have a relatively homogeneous microstructure without abnormal grains. The microstructure of the samples is consistent with the relative density data mentioned previously.

3.2. XRD analysis Figs. 6 and 7 demonstrate the phase analysis results by XRD. The major phases in the oxide additive samples are 3C and 4H polytypes, which are present in the starting SiC powders, therefore there is no new phase produced during the hot-pressing process. And as the content of the additives is small, XRD cannot detect them. In the nitrate salt additive samples S9N and S10S9N, XRD detects the diraction peaks of YAG (Al5Y3O12), which is the lowest eutectic phase between Al2O3 and Y2O3 and benecial to the fracture toughness of liquid phase sintered silicon carbide (LPSSC) ceramic [16,17]. Because of still lower additives content in

Fig. 3. The relative density comparison between the samples with oxide and nitrate salt additives.

Fig. 5. Secondary electron SEM images of the samples with nitrate additives (a) S9N, (b) S10S9N, (c) S10S6N, (d) S10S3N.

Fig. 4. Secondary electron SEM images of the samples with oxide additives (a) S9O, (b) S5S9O, (c) S10S9O, (d) S15S9O.

Fig. 6. XRD analysis of the samples with oxide additives: (a) starting SiC powder, (b) S9O, (c) S5S9O, (d) S10S9O, (e) S15S9O.

16

Z.H. Huang et al. / Ceramics International 29 (2003) 1317

S10S6N and S10S3N samples, the detected diraction peaks of these samples are the same as the starting SiC powders. 3.3. Elemental distribution analysis Fig. 8 shows the elemental distribution mapping in the samples S10S9O and S10S9N. Contrary to the results presented in Ref. [14], the distribution of elemental yttrium, aluminum and magnesium is more homogeneous in the samples with oxide additives. The dierence may be caused by the addition of nano-SiC particles. However, the authors also found many aggregates of the additives in the specimen of S9N. The only reason that can account for the phenomenon lies in the hot-pressing process. The detailed reasons for this dierence need further study.

Fig. 7. XRD analysis of the samples with nitrate additives (a) starting SiC powder, (b) S9N, (c) S10S9N, (d) S10S6N (e) S10S3N.

4. Conclusion Silicon carbide ceramic fabricated by hot-pressing using nanoscale SiC powder and nitrate salts as sintering additives had a ne, equiaxed microstructures with an average grain size less than 1 mm. The sintered samples with nitrate salt additives had higher relative density than the ones with oxide additives. The highest relative densities in the nitrate salt and oxide additive samples are 98.1 and 97.7%, respectively. XRD detected YAG phase, which is benecial to the fracture toughness of the ceramic, only in the nitrate salt additive samples. The electron probe analysis indicated that the elemental distribution of Y, Al, and Mg was more homogeneous in sintered samples with oxide additives.

References
[1] S.J. Guo, Theory of Powder Metallurgy, Metallurgical Industry Press, 1998, pp. 288291 (in Chinese). [2] S. Prochazka, The role of boron and carbon in the sintering of silicon carbide, in: P. Popper (Ed.), Special Ceramics, vol. 6, The British Ceramic Research Association, 1975, pp. 171181. [3] M. Omori, H. Takei, Pressureless sintering of SiC, J. Am. Ceram. Soc. 65 (6) (1982) C92. [4] M.A. Mulla, V.D. Krstic, Low-temperature pressureless sintering of b-silicon carbide with aluminum oxide and yttrium oxide additions, Am. Ceram. Soc. Bull. 70 (1991) 439443. [5] N.P. Padture, In situ-toughened silicon carbide, J. Am. Ceram. Soc. 77 (2) (1994) 519523. [6] S.K. Lee, Y.C. Kim, C.H. Kim, Microstructural development and mechanical properties of pressureless-sintered SiC with plate-like grains using Al2O3-Y2O3 additives, J. Mater. Sci. 29 (1994) 5321. [7] D. Foster, D.P. Thompson, The use of MgO as a densication aid for a-SiC, J. Eur. Ceram. Soc. 19 (1999) 28232831. [8] D.H. Kim, C.H. Kim, Toughening behavior of silicon carbide with additions of yttria and alumina, J. Am. Ceram. Soc. 73 (1990) 14311434.

Fig. 8. Elemental distribution mapping in the sintered samples: (a) Y in S10S9O, (b) Al in S10S9O, (c) Mg in S10S9O, (d) Y in S10S9N, (e) Al in S10S9N, (f) Mg in S10S9N.

Z.H. Huang et al. / Ceramics International 29 (2003) 1317 [9] M.A. Mulla, V.D. Krstic, Mechanical properties of b-SiC pressureless sintered with Al2O3 additions, Acta Mettall. Mater. 42 (1) (1994) 303308. [10] N.P. Padture, B.R. Lawn, Toughness properties of a silicon carbide with an in situ induced heterogeneous grain structure, J. Am. Ceram. Soc. 77 (1994) 25182522. [11] E. Liden, E. Carlstrom, L. Eklund, B. Nyberg, R. Carlsson, Homogeneous distribution of sintering additives in liquid-phase sintered silicon carbide, J. Am. Ceram. Soc. 78 (7) (1995) 1761 1768. [12] R. Vaben, A. Kaiser, J. Forster, H.P. Buchkremer, D. Stover, Densication of ultrane SiC powders, J. Mater. Sci. 31 (1996) 36233637. [13] J.-Y. Kim, T. Iseki, T. Yano, Pressureless sintering of dense

17

[14]

[15]

[16]

[17]

Si3N4 and Si3N4/SiC composites with nitrate additives, J. Am. Ceram. Soc. 79 (10) (1996) 27442746. D.-C. Park, T. Yano, T. Iseki, K. Urabe, Eect of nitrate salts as sintering additives during the ball-milling process of silicon nitride powders, J. Am. Ceram. Soc. 83 (12) (2000) 29672973. D. Sciti, A. Bellosi, Eect of additives on densication, microstructure and properties of liquid-phase sintered silicon carbide, J. Mater. Sci. 35 (2000) 38493855. J.H. She, K. Ueno, Densication behavior and mechanical properties of pressureless-sintered silicon carbide ceramics with alumina and yttria additions, Mater. Chem. Phys. 59 (1999) 139142. G. Magnani, G.L. Minoccari, L. Pilotti, Flexure strength and toughness of liquid phase sintered silicon carbide, Ceram. Int. 26 (2000) 495500.